By Lina Irawan 
The pursuit of AI dominance is a cornerstone of national strategies across Asia. Nearly every significant economy in the region has formally outlined a national AI masterplan, recognizing AI’s transformative potential for economic growth, innovation, and global competitiveness. Japan, for instance, recently unveiled an colossal 370 trillion yen ($2.3 trillion) budget, allocating more than a quarter of this sum specifically to artificial intelligence and semiconductor chip development over the next 15 years. This unprecedented investment signals a clear intent to be at the forefront of the AI revolution. Similarly, China continues its aggressive pursuit of AI leadership, while India positions itself as a global AI talent hub, and Singapore champions AI as a pillar of its Smart Nation vision. South Korea, too, is heavily investing in digital transformation and AI capabilities.
However, the ambition to lead in AI comes with an insatiable demand for power. Training advanced, "frontier" AI models necessitates concentrating immense computing power in a limited number of specialized locations, often requiring massive, dedicated data centers. Conversely, "inferencing"—the process of using trained AI models to make predictions or decisions—pushes the need for low-latency facilities into dense urban hubs, closer to end-users. The sheer scale of this energy requirement is staggering. Projections indicate that data center power demand across the Asia-Pacific region is set to surge by an estimated 165% between 2023 and 2030. This translates into hundreds of gigawatts of new electricity demand, a figure that many existing national grids are ill-equipped to handle.
Yet, much of the region’s impressive headline megawatt figures for data center capacity often fall into the category of "bragawatts"—grand announcements that look compelling on paper but are agonizingly slow to materialize into actual operational energy. This disconnect stems from a complex web of challenges inherent in infrastructure development.
The primary hurdle lies in the limitations of existing grid infrastructure. Despite rapid advancements in renewable energy generation, reliable power systems demand extensive and costly grid upgrades. Renewable sources, such as solar and wind, are typically sited far from major urban demand centers due to land availability and resource potential. Moreover, their intermittent nature—solar only when the sun shines, wind only when it blows—necessitates robust baseload power, sophisticated transmission networks, and substantial energy storage solutions to ensure grid stability and continuous supply. Without these critical investments in new transmission lines, smart grid technologies, and battery storage, the server racks of Asia’s burgeoning data centers will struggle to operate at full, consistent capacity. The International Energy Agency’s Southeast Asia Outlook starkly highlights this deficit: grid and storage investment in the region in 2023 was a mere $13 billion, dramatically short of the estimated $50 billion required annually until 2050 to meet projected demand and facilitate the energy transition.
Furthermore, in an era characterized by elevated fuel prices and pervasive energy insecurity, data centers often find themselves lower on the priority list for electricity supply. Policymakers and politicians face immense pressure to maintain essential services—keeping homes lit, hospitals running, and air-conditioning active—especially in regions prone to extreme weather. This prioritization can lead to power rationing or curtailment for less critical loads, including data centers, during peak demand periods or supply shortages.
The United States offers a stark cautionary tale. Despite its advanced infrastructure, up to half of all planned U.S. data center projects may not come online this year. The first three months of the year alone saw 75 data center projects, collectively valued at an estimated $130 billion, either blocked or significantly delayed due to mounting local opposition—a figure that astonishingly matches the total number of projects thwarted throughout all of 2023. Reasons for these delays are varied, ranging from "Not In My Backyard" (NIMBY) protests over noise pollution and aesthetic impact to grave concerns about the strain on local water resources, the overwhelming demand on existing power grids, and the lengthy interconnection queues for connecting to the transmission network. Supply chain bottlenecks, particularly for specialized equipment like large transformers, also contribute significantly to these delays.
Asia is already experiencing its own version of this predicament. The region managed to deliver only about 38% of its announced data center capacity in 2024, representing one of the widest "plan-to-delivery" gaps globally, according to a white paper co-authored with Oxford’s Smith School. This problem is particularly acute in nations like Malaysia and India, which have aggressively positioned themselves as digital infrastructure hubs. Malaysia’s state of Johor, for instance, a key data center hub, has imposed a ban on the construction of Tier 1 and 2 facilities due to escalating concerns over the immense strain these operations place on local water infrastructure, which is already struggling to meet residential and agricultural needs. Meanwhile, India’s ambitious goal to double its projected data center capacity by the end of the next financial year faces significant headwinds from severe grid delivery lags, compounded by challenges in land acquisition, right-of-way issues for transmission lines, and the sheer scale of investment required for grid modernization.
The pervasive AI-driven euphoria currently sweeping through financial markets is only exacerbating the disparity between what is promised and what is realistically feasible. This speculative fervor has tangible impacts on commodity markets. Copper prices, for example, have remained stubbornly high, buoyed by assumptions of surging data center construction demand, alongside demand from electric vehicles and renewable energy projects. This is despite the actual, slower pace of physical deployment. Similarly, the cost of critical equipment like power transformers has skyrocketed to two to three times their pre-2020 levels, as developers rush to secure scarce components, often locking in long lead times and inflated prices. If interconnection queues continue to stretch, mirroring the experiences in the U.S. and Europe, the profound mismatch between announced and delivered capacity could trigger a severe boom-bust cycle, reminiscent of the volatility observed in metals markets during the last decade. This could lead to stranded assets, cancelled projects, and a chilling effect on future investment.
In response to these burgeoning pressures, some Asian nations are proactively developing more robust regulatory frameworks. Singapore, Malaysia, and South Korea, for example, are now mandating that data center developers submit comprehensive plans for battery storage, demand curtailment management strategies, and thorough grid-impact assessments as part of their approval processes. While these measures are essential for sustainable development, they inevitably add layers of complexity and potentially slow down the pace of AI build-out in Asia. This measured approach contrasts sharply with the U.S., which despite its own challenges, is projected to forge ahead with an additional $4 trillion in data center construction planned through 2028. Every quarter that Asian operators delay in bridging this infrastructure gap translates into a quarter of lost compute capacity, talent migration, and capital deployment that could gravitate towards more readily available infrastructure elsewhere.
The fundamental solution lies in a more liberal and transparent approach to energy markets, specifically embracing distributed energy generation and facilitating robust electricity trading. This entails opening wholesale electricity markets to genuine price competition, which is crucial for attracting the substantial investment required to modernize grids and build new, reliable generation capacity. Furthermore, it would help diminish the region’s heavy reliance on imported oil and gas, bolstering energy security. However, achieving this requires significantly deeper and more transparent energy markets than currently exist across much of Asia.
Most Asian electricity systems are still rooted in a traditional, vertically integrated ecosystem. State-owned utilities typically act as single buyers, retail tariffs are often administratively set, and trading opportunities for third parties are severely limited. This structure stifles the generation of crucial future pricing signals, such as those derived from forward contracts, which are indispensable for long-term investment planning. Investors in mature markets, like those in Europe and the U.S., largely take such sophisticated trading layers for granted. Consequently, renewable energy investors in Asia face greater uncertainty regarding long-dated returns, as their projects are more exposed to the risks of government intervention during periods of high demand or supply shortages. This inherent uncertainty significantly increases the likelihood that critical grid projects stall due to a lack of sufficient investment—the very grid infrastructure that Asia’s data centers desperately need to connect to.
Despite these systemic challenges, pockets of progress are emerging. Japan’s power futures market, for instance, has rapidly become the fastest-growing electricity derivatives market globally, signaling a growing appetite for sophisticated risk management tools. India’s power exchange (IEX) has successfully launched day-ahead and term-ahead markets, providing clearer price signals and trading opportunities. Furthermore, the commercial flow and trading of electricity from Laos through Thailand and Malaysia to Singapore, which commenced in 2022, represents a groundbreaking step towards regional grid integration and cross-border energy trading, enhancing energy security and efficiency across Southeast Asia.
Companies like Marex are actively contributing to this evolution, boosting liquidity in Japan’s power derivatives markets. In New Zealand, Marex was recently selected to provide an over-the-counter (OTC) trading platform specifically designed to support the country’s standardized super-peak electricity contract. This initiative aims to deepen a local market that, much like many across Asia, has historically been fragmented and thinly traded.
These developments offer a glimpse into the future of deeper, more sophisticated Asian power markets. They envision a landscape where instruments and platforms empower generators, industrial users, and investors to hedge risks and price electricity with the same confidence and precision they apply to other established commodities like crude oil. For Asia to truly capitalize on its AI ambitions, its electricity markets must achieve a similar level of rigor, transparency, and liquidity as the global crude oil market. This requires proactive rather than reactive infrastructure development, driven by robust and transparent power markets that instill sufficient confidence in capital providers to invest ahead of demand, rather than perpetually playing catch-up. Ultimately, Asia’s economic prosperity in the AI era will hinge on its ability to build and sustain such resilient energy foundations.
